Mutations in p53 from different human cancers show mutagenic patterns that are presumed to provide insight into mutagenic mechanism. mC->T mutations are prominent in colon cancer, breast cancer and leukemias and are generally thought to be due to spontaneous deamination of m5C. In lung cancer there is a preponderance of G->T mutations in the non-transcribed strand of p53, which is reminiscent of mutagenesis by bulky agents, notably polycyclic aromatic hydrocarbons (PAHs) such as benzo[a]pyrene (BP). In liver cancer, which is associated with aflatoxin B1 (AFB1) exposure, there is an extraordinary hotspot (a G->T mutation at codon 249) in p53. The PI and collaborators have been involved in several significant studies, including one showing that slow DNA repair of UV damage correlated with skin cancer hotspots in p53, and hotspots for BP adduction (and slow DNA repair) correlated with lung cancer hotspots in p53. Such correlations are consistent with (but not proof of) a role for these etiological agents in carcinogenesis involving p53 mutations. In this proposal, the PI wishes to investigate these associations at a deeper level by studying the effect of adduction hotspots and differential repair directly on mutagenesis (using a variety of assays) by a number of agents. The focus will be on the mechanism of CpG mutational hotspots associated with the presence of m5C. The PI proposes three specific aims. (1) Sites for DNA adduct formation, repair and mutagenesis with BP and AFB1 will be determined in a single system (fibroblasts from the Big Blue mouse having a lacI mutational target). (2) UV-, BP- and AFB1-induced mutations in a yeast p53-containing vector system will be collected and compared to p53 mutations from human tumors. (3) The PI will investigate whether C->T hotspots at CpGs might be due to one of four endogenous mutagens, which might generate mutations preferentially at m5C residues.